Project Summary/Abstract: Approximately 75% of adults with serious mental illness (SMI) have a desire to work, but the level of unemployment for this population ranges from 85-90% (1-4). In addition to enjoying the benefit of increased income, research evidence indicates that persons with SMI who work also experience the additional benefits of increased self-esteem, satisfaction with finances, satisfaction with leisure and overall life satisfaction (5). Among the various types of vocational rehabilitation services currently available, supported employment (SE) is the most effective service for persons with SMI (13, 15, 16) and the VA has mandated that this evidence-based practice be offered as part of Compensated Work Therapy (CWT) Programs throughout the VA (21). Despite this mandate, there is a low rate of utilization of SE within the VA (22). Motivational interviewing (MI) is an evidence-based clinical practice with demonstrated effectiveness in producing desirable behavior change in the use of alcohol and drugs, as well other outcomes such as binge eating, smoking, treatment attendance, medication compliance and gambling (26- 34). Recently, Drebing (47), Glynn (45) and Mueller (46) have begun investigating the effectiveness of MI for increasing the rate of entry and length of participation n vocational rehabilitation programs among veterans with SMI. Mueller and her colleagues are currently conducting a clinical trial investigating the effectiveness of MI for increasing the rateof entry into SE and competitive employment outcomes. This two-year study analyzes archival data from the Mueller (46) study in which 79 unemployed veterans with serious mental illness (SMI) agreed to audiotaping and received either general supportive counseling (control) or Motivational Interviewing (experimental) in up to six sessions conducted over a period of six months. The purpose of the present study is to further analyze data from these research sessions to identify key ingredients of MI that are associated with the action of taking a step toward using SE services. Each MI research session is rated as having either a successful or unsuccessful outcome on the variable of taking a step toward supported SE utilization, which serves as the study dependent variable. The 206 audiotaped MI sessions conducted with participants in the experimental condition are coded using MISC 2.1 coding and analyzed using Generalized Estimating Equation (GEE) modeling to identify significant predictors of taking a step toward using SE. The independent variables are: a) therapist behaviors (spirit of MI and MI technical proficiency); and b) client behaviors (intensity and frequency of change talk and counter change talk). Covariates examined include age, race, mental health diagnosis, health status, work motivation, work importance, duration of unemployment, and type of disability benefits received. The study also analyzes sessions from 16 MI participants and 16 control participants using a grounded theory qualitative approach to identify emergent patterns and themes in the sessions. These findings are then combined with the quantitative findings to develop conceptual frameworks that explain differences between veterans with a successful outcome and those with an unsuccessful outcome. Study results are translated into recommendations regarding the components of MI that should receive greatest emphasis when training clinicians who will use MI with veterans who have SMI. The study also identifies additional factors influencing the taking of steps toward utilization of SE by veterans with SMI. New information obtained from this study will be useful for VA policy makers, administrators, clinicians, vocational rehabilitation specialists, and researchers interested in reducing the underutilization of Supported Employment, an evidence-based practice, and in improving employment outcomes that contribute to the overall recovery process for veterans with serious mental illness.